PDA

View Full Version : Shriner et al. (2014) Genome-wide genotype and sequence-based reconstruction of the 1



Drac II
03-12-14, 06:03
http://www.nature.com/srep/2014/140813/srep06055/full/srep06055.html#ref-link-42


We investigated ancestry of 3,528 modern humans from 163 samples. We identified 19 ancestral components, with 94.4% of individuals showing mixed ancestry. After using whole genome sequences to correct for ascertainment biases in genome-wide genotype data, we dated the oldest divergence event to 140,000 years ago. We detected an Out-of-Africa migration 100,000–87,000 years ago, leading to peoples of the Americas, east and north Asia, and Oceania, followed by another migration 61,000–44,000 years ago, leading to peoples of the Caucasus, Europe, the Middle East, and south Asia. We dated eight divergence events to 33,000–20,000 years ago, coincident with the Last Glacial Maximum. We refined understanding of the ancestry of several ethno-linguistic groups, including African Americans, Ethiopians, the Kalash, Latin Americans, Mozabites, Pygmies, and Uygurs, as well as the CEU sample. Ubiquity of mixed ancestry emphasizes the importance of accounting for ancestry in history, forensics, and health.

An interesting part of this genome-wide study is that it does not support the suggestion of Botigué et al. (2013) based on imprecise things like IBDs that there has been considerable gene-flow from Africa into southern Europe, particularly SW Europe (Portugal & Spain). The study did not find any significant African (either Northern or sub-Saharan) input in Spaniards:


Additionally, the PUR sample showed a significant amount of Berber ancestry, which likely did not derive from a Spanish parental population as none of the three Spanish samples (Spain_Basque, IBS (Iberian population in Spain), and Spain) showed significant amounts of Berber ancestry (Supplementary Fig. 4 and Supplementary Table 1)32.

The Spanish had no significant North African (labelled "Berber" in the study) or sub-Saharan African (the study divided sub-Saharan African into "Lowland East Cushitic", "Niger-Congo", "Nilo-Saharan", "Omotic" and "Pygmy") and the referred to Supplementary Table 1 simply considers it as zero. Northern Italians and Tuscans also had no significant African (either North African or sub-Saharan) and are also labelled as having zero in the table. The only European population sampled in the study that had some North African input were the Sardinians (6.1% "Berber") No European population sampled in the study was found to have any significant sub-Saharan African ancestry.

Pretty different, however, was the study's results regarding Middle Eastern (divided into two: "Arabian" and "Levantine-Caucasian") gene-flow into Europe. Here the study did confirm the suggestion of Botigué et al., as it found more significant amounts of geneflow from this area into Europe:


The Spanish and Italian samples showed southern and northern European ancestry with varying amounts of Levantine-Caucasian, Arabian, and Berber ancestry (Supplementary Fig. 5 and Supplementary Table 1). In contrast, the Basque samples showed only southern and northern European ancestry (Supplementary Fig. 5 and Supplementary Table 1), consistent with genetic isolation. Also, we detected more Arabian than Berber ancestry in Spain and Italy33. The oft-used CEU sample showed northern European, southern European, and Levantine-Caucasian ancestry, similar to the GBR (British in England and Scotland) and French samples (Supplementary Fig. 5 and Supplementary Table 1).
Note 33 refers to Botigué et al. and its IBD-based claims specifically.

The amount of Middle Eastern ancestry was higher in Romanians (5.9% "Arabian", 21.2% "Levantine-Caucasian"), then in North Italians & Tuscans (6.9 to 9% "Arabian", 17.2 to 21.1% "Levantine-Caucasian") and less in Spaniards (0 to 5% "Arabian", 6.8 to 9.6% "Levantine-Caucasian"), thus supporting the suggestion of Botigué et al. of a generally declining East-to-West gradient for this ancestry in Europe.

Fire Haired14
03-12-14, 06:53
I've been saying for a very long time that European's middle eastern ancestry is not all of Neolithic Stuttgart-like origin. Pretty much all Europeans, except Basque, show signs of modern middle eastern-type ancestry. Despite this pattern showing itself in every admixture test online, people for some reason ignore it.

It doesn't take a rocket scientist to notice there's post-Neolithic middle eastern ancestry in Europe, it just takes someone with eyes. F and D-statistics tell the same story.

This study and admixtures online, strongly suggests people in the Balkans and Italy have almost as much as 30% or more, middle eastern-type ancestry. That's pretty significant.

Still there's a very big divide between Europe and West Asia. Greeks are still probably something have around 70% European-specific ancestry(WHG+EHG+EEF), while Turks just east of them are probably well over 70% Middle eastern-specific. There's an even bigger divide between Volga Russians and Caucasians.

This means to me that Europe and West Asia should be considered two regions separated genetically, at the Caucasus mountains and Aegean sea.

We should probably use the term West Asia or Near east not Middle East, because north Africa and south-central Asia are probably very different.

polako
03-12-14, 08:15
This study is very basic.

A lot of the inferences are wrong or at the very least not very precise, probably because they're simply based on modern samples and ADMIXTURE, which is not a formal mixture test.

Botigué et al. was a lot more capable. I won't even try and compare this to Lazaridis et al.

Drac II
03-12-14, 08:40
This study is very basic.

A lot of the inferences are wrong or at the very least not very precise, probably because they're simply based on modern samples and ADMIXTURE, which is not a formal mixture test.

Botigué et al. was a lot more capable. I won't even try and compare this to Lazaridis et al.

IBDs can't even tell the direction of gene flow, and Lazaridis et al. themselves admitted the possibility of having overestimated admixture (at least in the case of the Spanish samples), so I would hardly call them better, at least not in this particular point.

Drac II
03-12-14, 08:45
I've been saying for a very long time that European's middle eastern ancestry is not all of Neolithic Stuttgart-like origin. Pretty much all Europeans, except Basque, show signs of modern middle eastern-type ancestry. Despite this pattern showing itself in every admixture test online, people for some reason ignore it.

It doesn't take a rocket scientist to notice there's post-Neolithic middle eastern ancestry in Europe, it just takes someone with eyes. F and D-statistics tell the same story.

This study and admixtures online, strongly suggests people in the Balkans and Italy have almost as much as 30% or more, middle eastern-type ancestry. That's pretty significant.

Still there's a very big divide between Europe and West Asia. Greeks are still probably something have around 70% European-specific ancestry(WHG+EHG+EEF), while Turks just east of them are probably well over 70% Middle eastern-specific. There's an even bigger divide between Volga Russians and Caucasians.

This means to me that Europe and West Asia should be considered two regions separated genetically, at the Caucasus mountains and Aegean sea.

We should probably use the term West Asia or Near east not Middle East, because north Africa and south-central Asia are probably very different.

Unfortunately, there were no truly "Balkan" samples in this study. The only sampled population for this study that can be partly considered so were the Romanians, and the study did find higher Middle Eastern there.

Goga
04-12-14, 03:52
I've been saying for a very long time that European's middle eastern ancestry is not all of Neolithic Stuttgart-like origin. Pretty much all Europeans, except Basque, show signs of modern middle eastern-type ancestry. Despite this pattern showing itself in every admixture test online, people for some reason ignore it.

It doesn't take a rocket scientist to notice there's post-Neolithic middle eastern ancestry in Europe, it just takes someone with eyes. F and D-statistics tell the same story.

This study and admixtures online, strongly suggests people in the Balkans and Italy have almost as much as 30% or more, middle eastern-type ancestry. That's pretty significant.

Still there's a very big divide between Europe and West Asia. Greeks are still probably something have around 70% European-specific ancestry(WHG+EHG+EEF), while Turks just east of them are probably well over 70% Middle eastern-specific. There's an even bigger divide between Volga Russians and Caucasians.

This means to me that Europe and West Asia should be considered two regions separated genetically, at the Caucasus mountains and Aegean sea.

We should probably use the term West Asia or Near east not Middle East, because north Africa and south-central Asia are probably very different.
Doesn't make any sense and I'll tell you why. Not everybody in Europe is the same and not everybody in West Asia is the same. Many ethnicities in Europe have different origins, like in West Asia many races have different origins. The distance between the folks from Caucasus is closer to North Eastern folks than people from Spain to Russians or Balthic people. The distance between Adygei and Russian or Baltic folks is smaller than distance between Spanish folks and the Russians of Lithuanians.On the other side, the distance between Arabs and Spanish people is much smaller than the distance between Spanish people and the Russians or Lithuanians. Spanish people are closer to the Arabs (Druze) than to the Russians. Because Russians and Spanish folks do NOT belong to the same race/species.


http://s12.postimg.org/tq815zob1/basque.jpg

Goga
04-12-14, 03:53
And as you can see on the map, the distance between West Asian (/Caucasian) Adygei and NorthEast European Russians is smaller than the distance between Adygei and the Arabs. So people in West Asia are also not the same and races of West Asia have different origins. But there's lots of race mixing going on, both within Europe and within West Asia. But as you can see there is a clear separation between NorthEastern Europe and SouthWestern Europe, like there is a clear separtaion bewtween NorthWestern Asian and SouthWest Asia...

Alan
04-12-14, 15:18
I've been saying for a very long time that European's middle eastern ancestry is not all of Neolithic Stuttgart-like origin. Pretty much all Europeans, except Basque, show signs of modern middle eastern-type ancestry. Despite this pattern showing itself in every admixture test online, people for some reason ignore it.the more modern Middle Eastern ancestry in Europeans started to come in during Bronze Age with Indo Europeans and went on with Etruscians, Pelasgians and the kind.
However "Arabian" in Europe, even though not exclusively, is probably mostly Neolithic farmer ancestry.

Although though it was explained by users like Angela a dozen times people seem to never understand that you can't and never should put ancient and more modern components in comparision as if these two stay in total contrast. You can't speak of EEF and West Asian or "European" and "none European" if both of these components share at least 70% of their ancestry, which stems from Early Near Eastern Farmers.

Because thats what it basically is. "West Asian" (proto herders) is 2/3 ENF and 1/3 ANE while EEF aka globe 13`s Mediterranean is 4/5 ENF and 1/5 WHG.
What differentiates Europeans from Western Asians is the additional WHG ancestry. And in this case the PCA plot fits perfect. If the difference was really 70% as someone might imply from your statement, than their would be an extremely huge gab between Greeks and Turks.

And as Goga pointed out, it's not like there is an invisible border between West Asians and Europeans which automatically makes any European population closer to any other European populations and any Western Asian to any other West Asian population. Still an Iberian is as far away from a Russian as an Armenian, Cypriot or Iranian. It's just that there is always a gab, or parallel evolution between Europeans and Western Asian populations because of the almost lack of WHG in one of the groups.

John Doe
04-12-14, 15:23
I've been saying for a very long time that European's middle eastern ancestry is not all of Neolithic Stuttgart-like origin. Pretty much all Europeans, except Basque, show signs of modern middle eastern-type ancestry. Despite this pattern showing itself in every admixture test online, people for some reason ignore it.

It doesn't take a rocket scientist to notice there's post-Neolithic middle eastern ancestry in Europe, it just takes someone with eyes. F and D-statistics tell the same story.

This study and admixtures online, strongly suggests people in the Balkans and Italy have almost as much as 30% or more, middle eastern-type ancestry. That's pretty significant.

Still there's a very big divide between Europe and West Asia. Greeks are still probably something have around 70% European-specific ancestry(WHG+EHG+EEF), while Turks just east of them are probably well over 70% Middle eastern-specific. There's an even bigger divide between Volga Russians and Caucasians.

This means to me that Europe and West Asia should be considered two regions separated genetically, at the Caucasus mountains and Aegean sea.

We should probably use the term West Asia or Near east not Middle East, because north Africa and south-central Asia are probably very different.

What about Ashkenazi Jews, Maltese and Sicilians? Where are they in this pattern? In the gap between Europe and west Asia?

Alan
04-12-14, 15:56
What about Ashkenazi Jews, Maltese and Sicilians? Where are they in this pattern? In the gap between Europe and west Asia?


Sicilians cluster with other Europeans but slightly closer to levantines as Greeks are. About Maltese I don't know.

But Sepharidc and Ashkenazi Jews fill the gab between Levantines (Druze) and Europeans. Ashkenazis tendings slightly more towards Europe than Sepharidic Jews. So it seems it is not entirely correct tos ay there is a gab anyways. At some point (Jews, Druze, Cypriots) Europeans and West Asians meet.

John Doe
04-12-14, 17:25
Sicilians cluster with other Europeans but slightly closer to levantines as Greeks are. About Maltese I don't know.

But Sepharidc and Ashkenazi Jews fill the gab between Levantines (Druze) Europeans. Ashkenazis tendings slightly more towards Europe than Sepharidic Jews. So it seems it is not entirely correct tos ay there is a gab anyways. At some point (Jews, Druze, Cypriots) Europeans and West Asians meet.

PCAs in Lazaridis et al:

http://postimg.org/image/ll1tqocod/
http://postimg.org/image/8wpeopdff/

Alan
04-12-14, 17:30
PCAs in Lazaridis et al:

http://postimg.org/image/ll1tqocod/
http://postimg.org/image/8wpeopdff/
It's hard to spot the populations on those two images. It's also important to note on what purpose and dimension these PCA plots are made. For example the first one seems to compare modern populations to EEF and ANE and not modern populations to each other.

John Doe
04-12-14, 17:33
It's hard to spot the populations.
I apologise, that was the closest I could get.

Alan
04-12-14, 18:23
I apologise, that was the closest I could get.

Well to be honest I can spot the populations but it's hard to tell on what context the map was made.

John Doe
04-12-14, 18:52
Well to be honest I can spot the populations but it's hard to tell on what context the map was made.
Both the first and second one compare modern vs. modern and modern vs. ancient.

Hauteville
04-12-14, 18:59
Maltese are just sicilians and deep south italians genetically but with more north african admixture they however plot not distant to sicilian and calabrese. They are in the italic genetic continuum.
http://i57.tinypic.com/mlla9.jpg

John Doe
04-12-14, 20:40
Maltese are just sicilians and deep south italians genetically but with more north african admixture they however plot not distant to sicilian and calabrese. They are in the italic genetic continuum.
http://i57.tinypic.com/mlla9.jpg

Maltese plot pretty much exactly where Sicilians do, but so do Ashkenazis, according to this logic AJs are also Italic, and so are Greeks who also plot in the same place.

Hauteville
04-12-14, 20:56
You did not understand what i meant. If for someone maltese are different to some italians they surely don't plot near sicilians. Unfortunately there is only one sample from continental south from Calabria and the comparison is not possible to do. However sicilians are much more close to tuscans and north italians than middle east and the same for greeks with balkans.

Angela
04-12-14, 21:45
the more modern Middle Eastern ancestry in Europeans started to come in during Bronze Age with Indo Europeans and went on with Etruscians, Pelasgians and the kind.
However "Arabian" in Europe, even though not exclusively, is probably mostly Neolithic farmer ancestry.

Although though it was explained by users like Angela a dozen times people seem to never understand that you can't and never should put ancient and more modern components in comparision as if these two stay in total contrast. You can't speak of EEF and West Asian or "European" and "none European" if both of these components share at least 70% of their ancestry, which stems from Early Near Eastern Farmers.

Because thats what it basically is. "West Asian" (proto herders) is 2/3 ENF and 1/3 ANE while EEF aka globe 13`s Mediterranean is 4/5 ENF and 1/5 WHG.
What differentiates Europeans from Western Asians is the additional WHG ancestry. And in this case the PCA plot fits perfect. If the difference was really 70% as someone might imply from your statement, than their would be an extremely huge gab between Greeks and Turks.

And as Goga pointed out, it's not like there is an invisible border between West Asians and Europeans which automatically makes any European population closer to any other European populations and any Western Asian to any other West Asian population. Still an Iberian is as far away from a Russian as an Armenian, Cypriot or Iranian. It's just that there is always a gab, or parallel evolution between Europeans and Western Asian populations because of the almost lack of WHG in one of the groups.

Well, thank God for Alan! At least somebody gets it.http://cdn.eupedia.com/forum/images/smilies/main/smile.gif Or, to put it another way, as I said on a recent thread, if you compare EEF ancestry to modern Near Easterners, just take out the WHG carried by the EEF people (We don't know what percentage to ascribe to it actually, although we could maybe use an estimate of 20%? That might be too high, but whatever.), add a big chunk of ANE (which also went into northern Europe) and then for the areas south of, perhaps, Syria, add anywhere from 4 up to 7% probably mostly recent SSA. Now, that's a fast and dirty analysis but the essentials are correct, I think.

I'm at a loss as to why there is all this confusion about this matter. It just takes a little logic. To look at it yet another way, if the Yamnaya people turn out to indeed be half ancient Karelian like, thus having a decidedly "eastern, north Eurasian" strain, and half modern "Armenian like", would they be "European" according to these templates? It just gets nonsensical when people try to apply modern geographic, political. and cultural divisions to ancient Dna.

In terms of the paper itself, I did finally read it, although belatedly. I initially put it down after reading that it was based on "19,372 autosomal single nucleotide polymorphisms". It's just marginally better than using AIMs. Then add to that the fact that they base all their conclusions on "Admixture" programs of modern populations. Somebody should send them a PM with all of the more recent studies using ancient DNA.

As to the Botigue et al study, it's true that as Dienekes pointed out you can't take their percentages as to IBD sharing as gospel. The figures and graphs in that paper may not be precise because the authors of that particular study did not analyze the geographic origin of those particular IBD segments. However, that there was gene flow, for example, from North Africa across the Strait of Gibraltar is undeniable. All you have to do is trace the occurrence of E-M81. It also exists in Sicily, and some other parts of Italy, (as we discussed on another thread) although not at those levels. We also don't currently have enough subclade resolution or certainty about mutation rates to know precisely when it arrived in either place, although recent papers are providing that kind of information at least in terms of North Africa. The mtDna paper on U6 was very good, I thought. We need that kind of analysis of uniparental markers in Europe. Unfortunately, all the money and research seems to be going into R1b and R1a, which is understandable, of course, given the numbers in Europe. The "I" clades have Nordvelt, of course. We need someone like him for "G" and "E" and the other more minor clades. As to the SSA segments, could anyone honestly argue that they moved from Europe to Africa?

See:
http://www.eupedia.com/images/content/Haplogroup-E-M81.gif

See also this paper about Moroccan E-M81 showing a decreasing gradient of E-M81 as one moves from south to north. It sounds as if there could be lots of pertinent information in there, but it's behind a paywall. If someone has institutional access, it would be great to get some nuggets from it.
http://www.ncbi.nlm.nih.gov/pubmed/25397701?dopt=Abstract

Also see Dienekes' analysis of Botigue et al:
http://dienekes.blogspot.com/2013/06/ibd-sharing-between-iberians-and-north.html

John Doe
04-12-14, 22:04
However sicilians are much more close to tuscans and north italians than middle east and the same for greeks with balkans.
If that's the case then AJs are also closer to Europe than the middle east.

Alan
04-12-14, 22:31
6897


On this map we see perfectly what Angela and I described. That it is the WHG admixture which gives the Europeans a parallel evolution to West Asians. In West Asians replace 30% of their total genome with WHG. And the West Asian line will almost overlap with the European line. North Caucasians will fall into North/Northeastern cluster and Levantines over the Spanish/Sardinian cluster. Which perfectly visualizes that Sardinians are ancient Levantines + additional WHG.

Hauteville
04-12-14, 23:08
If that's the case then AJs are also closer to Europe than the middle east.
Italy and Balkans are Europe so yes ;)

Greying Wanderer
05-12-14, 06:40
@Angela


if you compare EEF ancestry to modern Near Easterners, just take out the WHG carried by the EEF people (We don't know what percentage to ascribe to it actually, although we could maybe use an estimate of 20%? That might be too high, but whatever.)

I thought EEF was defined as where the WHG component within EEF was below 50% so the WHG part of EEF could range from 0% to 49%.

(I may be misremembering though.)

edit
(Although I think having two components where one (WHG) is both one of the components itself and also a component within the other component seems unnecessarily confusing. Wouldn't it be better to say 70% WHG / 30% Basal or 40% WHG and 60% Basal instead of percentages of WHG and EEF.)

Alan
05-12-14, 17:19
@Angela



I thought EEF was defined as where the WHG component within EEF was below 50% so the WHG part of EEF could range from 0% to 49%.

(I may be misremembering though.)

edit
(Although I think having two components where one (WHG) is both one of the components itself and also a component within the other component seems unnecessarily confusing. Wouldn't it be better to say 70% WHG / 30% Basal or 40% WHG and 60% Basal instead of percentages of WHG and EEF.)

In the Lazaridis paper it's 49% at max and 2% at minimum. But both 49% and 2% are the dead end and the most unlikely scenario. So 25% would be the median. But the paper also says the more African in Bedouins and WHG in Sardinians we presume, the smaller the percentage is. And since we know both is the case it is even much more likely that WHG admixture in EEF is below 25%.

Keep in mind Early Near Eastern Farmer != Basal Eurasian. Not all H&G admixture in EEF is WHG. You can bet that ENF was already H&G admixed in the Near East. ENF on itself was already a Basal Eurasian+H&G mixed component. Just the additional 20% Western H&G which they catched up, most probably in the Balkans. Is what differentiates EEF from ENF.

Angela
05-12-14, 18:44
In the Lazaridis paper it's 49% at max and 2% at minimum. But both 49% and 2% are the dead end and the most unlikely scenario. So 25% would be the median. But the paper also says the more African in Bedouins and WHG in Sardinians we presume, the smaller the percentage is. And since we know both is the case it is even much more likely that WHG admixture in EEF is below 25%.

Keep in mind Early Near Eastern Farmer != Basal Eurasian. Not all H&G admixture in EEF is WHG. You can bet that ENF was already H&G admixed in the Near East. ENF on itself was already a Basal Eurasian+H&G mixed component. Just the additional 20% Western H&G which they catched up, most probably in the Balkans. Is what differentiates EEF from ENF.

It's a distinction without a difference, Alan, but I think you have a typo there. I believe it's 2 to 45%.

Just so that it's written down somewhere since people seem to forget it, here is the quote from the Supplement, Section 13:

"The amount of Near Eastern admixture estimated for Stuttgart can be seen in Table S13.2 and ranges between 55-100. There are reasons to doubt both the lower estimates (near 55%), since ALDER provides only a lower bound on African ancestry, but also the higher estimates (near 100%) since there is direct evidence that Stuttgart has European huntergatherer ancestry."

I think they mention the 2% somewhere else, but I couldn't quickly find it. So, 2-45%, with an average of 23%?Like you, I think it's lower, however, for the reasons you gave.


Of course, Lazaridis et al is at pains to point out that we won't know anything with precision until we get an ancient farmer sample from the Near East, and so these numbers may change as we hopefully get more ancient Dna.

Greying Wanderer
07-12-14, 06:22
"The amount of Near Eastern admixture estimated for Stuttgart can be seen in Table S13.2 and ranges between 55-100. There are reasons to doubt both the lower estimates (near 55%), since ALDER provides only a lower bound on African ancestry, but also the higher estimates (near 100%) since there is direct evidence that Stuttgart has European huntergatherer ancestry."

I'm not disputing Stuttgart but the overall sequence:

East Asia
Farmer expansion -> displacement of HG dna -> The End, roll credits

Western version
Farmer expansion -> initial displacement of HG dna -> IE + cousins of HG -> various other stuff etc

with the overall effect being the farmer expansion was stalled partway through.

(i say stalled rather than stopped as i think the farmer dna may have bounced back over time but the "over time" bit is where the big differences with east asia lie imo.)

Angela
07-12-14, 06:34
I'm not disputing Stuttgart but the overall sequence:

East Asia
Farmer expansion -> displacement of HG dna -> The End, roll credits

Western version
Farmer expansion -> initial displacement of HG dna -> IE + cousins of HG -> various other stuff etc

with the overall effect being the farmer expansion was stalled partway through.

(i say stalled rather than stopped as i think the farmer dna may have bounced back over time but the "over time" bit is where the big differences with east asia lie imo.)

You may have forgotten, but I found the quote from Lazaridis in response to this statement from you:

I thought EEF was defined as where the WHG component within EEF was below 50% so the WHG part of EEF could range from 0% to 49%.

joeyc
12-12-14, 17:00
This paper only uses 19.000 SNPs. Low SNPs Admixture analysis are known to overextimate admixture. I think that it's not very correct.

For example Puerto Ricans score 30% of undefined admixture.

Sigfrido
16-10-15, 09:55
This paper only uses 19.000 SNPs. Low SNPs Admixture analysis are known to overextimate admixture. I think that it's not very correct.For example Puerto Ricans score 30% of undefined admixture.Indeed one of the crappiest and dumbest paper published so far. Most Europeans get a lot of undefined admixture.The authors were a bunch of baboons.